Abstract
Process Mining offers a powerful framework for uncovering, analyzing, andoptimizing real-world business processes. Petri nets provide a versatile meansof modeling process behavior. However, traditional methods often struggle toeffectively compare complex Petri nets, hindering their potential for processenhancement. To address this challenge, we introduce PetriNet2Vec, anunsupervised methodology inspired by Doc2Vec. This approach converts Petri netsinto embedding vectors, facilitating the comparison, clustering, andclassification of process models. We validated our approach using the PDCDataset, comprising 96 diverse Petri net models. The results demonstrate thatPetriNet2Vec effectively captures the structural properties of process models,enabling accurate process classification and efficient process retrieval.Specifically, our findings highlight the utility of the learned embeddings intwo key downstream tasks: process classification and process retrieval. Inprocess classification, the embeddings allowed for accurate categorization ofprocess models based on their structural properties. In process retrieval, theembeddings enabled efficient retrieval of similar process models using cosinedistance. These results demonstrate the potential of PetriNet2Vec tosignificantly enhance process mining capabilities.